Reverting an old snapshot on a production volume without a full sweep

Abstract

It may be beneficial to revert from the production volume V to the production snapshot S. Traditional approaches required a full sweep of production volume data when reverting to a snapshot (i.e., reinitialize all data, mark all data as dirty and start replicating to update the replication volume V′ to what the production volume V stores (i.e., the former production snapshot S). However, example embodiments of the present invention provide for reverting from a production volume to a snapshot without requiring a full sweep of data in the production volume. Rather, example embodiments of the present invention stop replication of the production volume, notify a splitter of dirty location in the snapshot, synchronize the dirty locations with the replication volume and resume replication to the snapshot, thereby performing a minimal initialization.

What is claimed is: 
     
       1. A method for reverting from a production volume to a snapshot in a replication environment comprising a production site including the production volume, the snapshot, and a splitter, the method comprising:
 stopping replication of the production volume; 
 setting a revert flag in the splitter; 
 reverting the production volume to the snapshot by notifying the splitter of dirty locations in the snapshot and tracking metadata of the dirty locations in the splitter; 
 clearing the revert flag to notify the splitter that the revert is completed; and 
 resuming replication of the production volume. 
 
     
     
       2. The method of  claim 1  wherein the replication environment further comprises a replication site including a replication volume and wherein resuming replication of the production volume further comprises:
 reading the metadata from the splitter; and 
 synchronizing the dirty locations with the replication volume. 
 
     
     
       3. The method of  claim 1  wherein stopping replication of the production volume comprises transitioning the splitter from a split mode to a marking on host mode. 
     
     
       4. The method of  claim 3  wherein resuming replication of the production volume comprises:
 transitioning the splitter from the marking on host mode to the split mode once the revert flag is cleared; 
 receiving dirty location metadata from the splitter for each dirty location in the snapshot; and 
 synchronizing the dirty locations to the replication volume. 
 
     
     
       5. The method of  claim 1  wherein reverting the production volume to the snapshot comprises:
 moving the data of the snapshot to the production volume; and 
 deleting the snapshot data and the old production volume data. 
 
     
     
       6. An apparatus for reverting from a production volume to a snapshot in a replication environment comprising a production site including the production volume, the snapshot, and a splitter, the apparatus comprising:
 a controller configured to stop replication of the production volume,
 set a revert flag in the splitter, 
 clear the revert flag to notify the splitter that the revert is completed, and 
 resume replication of the production volume after reverting the production volume to the snapshot; and 
 a manager configured to revert the production volume to the snapshot by notifying the splitter of dirty locations in the snapshot and tracking metadata of the dirty locations in the splitter. 
 
 
     
     
       7. The apparatus of  claim 6  wherein the replication environment further comprises a replication site including a replication volume and wherein the controller is further configured to read the metadata from the splitter and synchronize the dirty locations with the replication volume. 
     
     
       8. The apparatus of  claim 6  wherein the controller is further configured to transition the splitter from a split mode to a marking on host mode. 
     
     
       9. The apparatus of  claim 8  wherein the controller is further configured to transition the splitter from the marking on host mode to the split mode, once the revert flag is cleared, receive dirty location metadata from the splitter for each dirty location in the snapshot and synchronize the dirty locations to the replication volume. 
     
     
       10. The apparatus of  claim 6  wherein the manager is further configured to move the data of the snapshot to the production volume and delete the snapshot data and the old production volume data. 
     
     
       11. A computer-program product including a non-transitory computer-readable storage medium encoded with computer-program code that, when executed on a processor of a computer, cause the computer to revert from a production volume to a snapshot in a replication environment comprising a production site including the production volume, the snapshot, and a splitter, the computer-program code comprising:
 computer-program code for stopping replication of the production volume; 
 computer-program code for setting a revert flag in the splitter; 
 computer-program code for reverting the production volume to the snapshot notifying the splitter of dirty locations in the snapshot and tracking metadata of the dirty locations in the splitter; 
 computer-program code for clearing the revert flag to notify the splitter that the revert is completed; and 
 computer-program code for resuming replication to the production volume. 
 
     
     
       12. The computer-program product of  claim 11  wherein the replication environment further comprises a replication site including a replication volume and wherein computer-program code for resuming replication to the production volume further comprises:
 computer-program code for reading the metadata from the splitter; and 
 computer-program code for synchronizing the dirty locations with the replication volume. 
 
     
     
       13. The computer-program product of  claim 11  wherein computer-program code for stopping replication of the production volume comprises computer-program code for transitioning the splitter from a split mode to a marking on host mode and wherein computer-program code for resuming replication of the production volume to the snapshot comprises:
 computer-program code for transitioning the splitter from the marking on host mode to the split mode once the revert flag is cleared; 
 computer-program code for receiving dirty location metadata from the splitter for each dirty location in the snapshot; and 
 synchronizing the dirty locations to the replication volume. 
 
     
     
       14. The computer-program product of  claim 11  wherein computer-program code for reverting the production volume to the snapshot comprises:
 computer-program code for moving the data of the snapshot to the production volume; and 
 computer-program code for deleting the snapshot data and the old production volume data.